DESIGN AND IMPLEMENTATION OF DRIVER DROWSINESS DETECTION AND WARNING SYSTEM BASED ON STANDARD DEVIATION OF LATERAL POSITION
Driving in a drowsy state causes decreased concentration in the driver and can cause traffic accidents. One of the things that can indicate a sleepy driver is that the vehicle is moving away from the proper lane. In addition to internal factors, external factors such as roads with monotonous charact...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/64057 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Driving in a drowsy state causes decreased concentration in the driver and can cause traffic accidents. One of the things that can indicate a sleepy driver is that the vehicle is moving away from the proper lane. In addition to internal factors, external factors such as roads with monotonous characteristics and low traffic density also can reduce driver alertness because it only slightly stimulates the driver's movement. Drowsiness detection and warning systems are developed for drivers to reduce the potential of traffic accidents due to drowsy drivers. The proposed design of the system is a hybrid system that detects drowsiness from the driver's behavioral and vehicle-based measures. One of the detection systems designed is the drowsiness detection based on the Standard Deviation of Lateral Position (SDLP) value. OpenCV library and Python programming language are used in processing road images and calculating SDLP values. The value of SDLP is estimated from the vehicle's position against the lane markings. The process begins with a preprocessing step consisting of several digital image processing techniques such as thresholding, grayscaling, and Canny edge detection. Then, the pixel position of the lane markings will be searched to calculate the SDLP value. The SDLP value that has been obtained will be compared to the estimated SDLP value of Karolinska Sleepiness Scale 7 and the warning alarm will ring if the SDLP value is greater than or equal to 0.277 meters.
Tests were conducted on several road images with daytime and night conditions. The accuracy of the Standard Deviation of Lateral Position value calculation reaches 96.31% on the daytime road images and 90.25% on the night road images. The warning alarm successfully rang when the SDLP value was greater than or equal to 0.277 meters. |
|---|